How enterprises can navigate ethics and responsibility of generative AI

In a few short months, generative AI has become a very hot topic. Looking beyond the hype, generative AI is a groundbreaking technology, enabling novel capabilities as it moves rapidly into the enterprise world. 

According to a CRM survey, 67% of IT leaders are prioritizing generative AI for their business within the next year and a half—despite looming concerns about generative AI ethics and responsibility. And 80% of those who think generative AI is “overhyped” still believe the technology will improve customer support, reduce workloads and boost organizational efficiencies.

In the enterprise world, generative AI has arrived (discussed in my previous CIO.com article about enterprises putting generative AI to work here).

Preserving trust

As enterprises race to adopt generative AI and begin to realize its benefits, there is a simultaneous mandate in play. Organizations must proactively mitigate generative AI’s inherent risks, in areas such as ethics, bias, transparency, privacy and regulatory requirements.

Fostering a responsible approach to generative AI implementations enables organizations to preserve trust with customers, employees and stakeholders. Trust is the currency of business. Without it, brands can be damaged as revenues wane and employees leave. And once breached, trust is difficult to regain. 

That’s why preserving trust—before it is broken—is so essential. Here are ways to proactively preserve trust in generative AI implementations.

Mitigating bias and unfairness

Achieving fairness and mitigating bias are essential aspects of responsible AI deployment. Bias can be unintentionally introduced from the AI training data, algorithm and use case. Picture a global retail company using generative AI to personalize promotional offers for customers. The retailer must prevent biased outcomes like offering discounts to specific demographic groups only. 

To do that, the retailer must create diverse and representative data sets, employing advanced techniques for bias detection and mitigation and adopting inclusive design practices. Ongoing, the continuous monitoring and evaluation of AI systems will ensure fairness is maintained throughout their lifecycle.

Establishing transparency and explainability

In addition to mitigating bias and unfairness, transparency and explainability in AI models are vital for establishing trust and ensuring accountability. Consider an insurance company using generative AI to forecast claim amounts for its policyholders. When the policyholders receive the claim amounts, the insurer needs to be able to explain the reasoning behind how they were estimated, making transparency and explainability fundamental.

Due to the complex nature of AI algorithms, achieving explainability, while essential, can be challenging. 

However, organizations can invest in explainable AI techniques (e.g., data visualization or decision tree), provide thorough documentation and foster a culture of open communication about the AI decision-making processes. 

These efforts help demystify the inner workings of AI systems and promote a more responsible, transparent approach to AI deployment.

Safeguarding privacy

Privacy is another key consideration for responsible AI implementation. Imagine a healthcare organization leveraging generative AI to predict patient outcomes based on electronic health records. Protecting the privacy of individuals is a must-have, top priority. Generative AI can inadvertently reveal sensitive information or generate synthetic data resembling real individuals. 

To address privacy concerns, businesses can implement best practices like data anonymization, encryption and privacy-preserving AI techniques, such as differential privacy. Concurrently, organizations must remain compliant with data protection regulations such as the General Data Protection Regulation (GDPR) and the Health Insurance Portability and Accountability Act (HIPAA).

Complying with regulatory requirements

Finally, the evolving regulatory landscape for AI technologies demands a robust governance framework that guides ethical and responsible AI deployment. 

Organizations can refer to resources like the European Union’s Ethics Guidelines for Trustworthy AI or the Organisation for Economic Cooperation and Development (OECD) AI Principles to help define AI policies and principles. Establishing cross-functional AI ethics committees and developing processes for monitoring and auditing AI systems help organizations stay ahead of regulatory changes. By adapting to changes in regulations and proactively addressing potential risks, organizations can demonstrate their commitment to responsible AI practices. 

Responsible AI deployment

At Dell Technologies, we have articulated our ethical AI principles here. We know that responsible AI use plays a crucial role in an enterprise’s successful adoption of generative AI. To realize the extraordinary potential of generative AI, organizations must continuously improve and adapt their practices and address evolving ethical challenges like bias, fairness, explainability, transparency, privacy preservation and governance. 

Read about enterprise use cases for generative AI in this CIO.com article.

***

Dell Technologies. To help organizations move forward, Dell Technologies is powering the enterprise generative AI journey. With best-in-class IT infrastructure and solutions to run generative AI workloads and advisory and support services that roadmap generative AI initiatives, Dell is enabling organizations to boost their digital transformation and accelerate intelligent outcomes. 

Intel. The compute required for generative AI models has put a spotlight on performance, cost and energy efficiency as top concerns for enterprises today. Intel’s commitment to the democratization of AI and sustainability will enable broader access to the benefits of AI technology, including generative AI, via an open ecosystem. Intel’s AI hardware accelerators, including new built-in accelerators, provide performance and performance per watt gains to address the escalating performance, price and sustainability needs of generative AI.



Source link